IEEE VIS Publication Dataset

Palantir is an analytic platform currently used worldwide by both governmental and financial analysts. This paper provides a brief overview of the platform, examines our 2009 IEEE VAST Challenge submission, and highlights several key analytic and visualization features we used in our analysis.

Do court cases differ from place to place? What kind of picture do we get by looking at a country's collection of law cases? We introduce parallel tag clouds: a new way to visualize differences amongst facets of very large metadata-rich text corpora. We have pointed parallel tag clouds at a collection of over 600,000 US Circuit Court decisions spanning a period of 50 years and have discovered regional as well as linguistic differences between courts. The visualization technique combines graphical elements from parallel coordinates and traditional tag clouds to provide rich overviews of a document collection while acting as an entry point for exploration of individual texts. We augment basic parallel tag clouds with a details-in-context display and an option to visualize changes over a second facet of the data, such as time. We also address text mining challenges such as selecting the best words to visualize, and how to do so in reasonable time periods to maintain interactivity.

IceXplorer is a tool for analyzing variations in ice cover on Lake Erie. It enhances the data and pre-packaged analysis currently available in the great lakes ice atlas and serves as an example of a small, focused application where simple but carefully-chosen visualizations, interaction techniques, and automated data analysis are combined to create an effective tool for advancing scientific research.

Many well-known time series prediction methods have been used daily by analysts making decisions. To reach a good prediction, we introduce several new visual analysis techniques of smoothing, multi-scaling, and weighted average with the involvement of human expert knowledge. We combine them into a well-fitted method to perform prediction. We have applied this approach to predict resource consumption in data center for next day planning.

In modern process industry, it is often difficult to analyze a manufacture process due to its numerous time-series data. Analysts wish to not only interpret the evolution of data over time in a working procedure, but also examine the changes in the whole production process through time. To meet such analytic requirements, we have developed ProcessLine, an interactive visualization tool for a large amount of time-series data in process industry. The data are displayed in a fisheye timeline. ProcessLine provides good overviews for the whole production process and details for the focused working procedure. A preliminary user study using beer industry production data has shown that the tool is effective.

Professional cyber analysts were observed as they attempted to solve the VAST 2009 Traffic Mini Challenge using basic visualization tools and a large, high-resolution display. We discuss some of the lessons we learned about how analysts actually work and potential roles for visualization and large, high-resolution displays.

The increasing availability of motion sensors and video cameras in living spaces has made possible the analysis of motion patterns and collective behavior in a number of situations. The visualization of this movement data, however, remains a challenge. Although maintaining the actual layout of the data space is often desirable, direct visualization of movement traces becomes cluttered and confusing as the spatial distribution of traces may be disparate and uneven. We present proximity-based visualization as a novel approach to the visualization of movement traces in an abstract space rather than the given spatial layout. This abstract space is obtained by considering proximity data, which is computed as the distance between entities and some number of important locations. These important locations can range from a single fixed point, to a moving point, several points, or even the proximities between the entities themselves. This creates a continuum of proximity spaces, ranging from the fixed absolute reference frame to completely relative reference frames. By combining these abstracted views with the concrete spatial views, we provide a way to mentally map the abstract spaces back to the real space. We demonstrate the effectiveness of this approach, and its applicability to visual analytics problems such as hazard prevention, migration patterns, and behavioral studies.

The classic TileBars paradigm has been used to show distribution information of query terms in full-text documents. However, when the number of query terms becomes large, it is not an easy task for users to comprehend their distribution within certain parts of a document. In this paper, we present a novel approach to improve the visual presentation of TileBars, in which barycenter heuristic for bigraph crossing minimization is used to reorder TileBars elements. The reordered TileBars can be demonstrated to provide users with better focus and navigation while exploring text documents.

We present our visualization systems and findings for the badge and network traffic as well as the social network and geospatial challenges of the 2009 VAST contest. The summary starts by presenting an overview of our time series encoding of badge information and network traffic. Our findings suggest that employee 30 may be of interest. In the second part of the paper, we describe our system for finding subgraphs in the social network subject to degree constraints. Subsequently, we present our most likely candidate network which is similar to scenario B.

We present a new application, SpRay, designed for the visual exploration of gene expression data. It is based on an extension and adaption of parallel coordinates to support the visual exploration of large and high-dimensional datasets. In particular, we investigate the visual analysis of gene expression data as generated by micro-array experiments; We combine refined visual exploration with statistical methods to a visual analytics approach that proved to be particularly successful in this application domain. We will demonstrate the usefulness on several multidimensional gene expression datasets from different bioinformatics applications.

Our visualization tool for the VAST 2009 traffic mini challenge, Timeliner, visualizes badge and network traffic data together in a single timeline. The two views of per-employee and per-day with various filtering interactions enable users to analyze easily employees activities at a particular moment of interest as well as their general daily patterns. Using Timeliner, we present several hypotheses for the task at hand and their validation processes, which reveals various aspects of the data.

In this paper, we discuss dimension reduction methods for 2D visualization of high dimensional clustered data. We propose a two-stage framework for visualizing such data based on dimension reduction methods. In the first stage, we obtain the reduced dimensional data by applying a supervised dimension reduction method such as linear discriminant analysis which preserves the original cluster structure in terms of its criteria. The resulting optimal reduced dimension depends on the optimization criteria and is often larger than 2. In the second stage, the dimension is further reduced to 2 for visualization purposes by another dimension reduction method such as principal component analysis. The role of the second-stage is to minimize the loss of information due to reducing the dimension all the way to 2. Using this framework, we propose several two-stage methods, and present their theoretical characteristics as well as experimental comparisons on both artificial and real-world text data sets.

Gene mapping is a statistical method used to localize human disease genes to particular regions of the human genome. When performing such analysis, a genetic likelihood space is generated and sampled, which results in a multidimensional scalar field. Researchers are interested in exploring this likelihood space through the use of visualization. Previous efforts at visualizing this space, though, were slow and cumbersome, only showing a small portion of the space at a time, thus requiring the user to keep a mental picture of several views. We have developed a new technique that displays much more data at once by projecting the multidimensional data into several 2D plots. One plot is created for each parameter that shows the change along that parameter. A radial projection is used to create another plot that provides an overview of the high dimensional surface from the perspective of a single point. Linking and brushing between all the plots are used to determine relationships between parameters. We demonstrate our techniques on real world autism data, showing how to visually examine features of the high dimensional space.

The 4th VAST Challenge centered on a cyber analytics scenario and offered three mini-challenges with datasets of badge and network traffic data, a social network including geospatial information, and security video. Teams could also enter the Grand challenge which combined all three datasets. In this paper, we summarize the dataset, the overall scenario and the questions asked in the challenges. We describe the judging process and new infrastructure developed to manage the submissions and compute accuracy measures in the social network mini challenge. We received 49 entries from 30 teams, and gave 23 different awards to a total of 16 teams.

As a solution to the VAST 2009 Traffic Mini Visualization Challenge, we built the Badge and Network Traffic (BNT) tool to create animations of the events taking place in the embassy. Using the embassy layout, the prox-card and web-access entries and their time-stamps, we animated color-based flagging of events. The BNT tool highlights logical anomalies occuring in the badge and network traffic data with color-coded alerts. Prior to the animated visualization, the tool analyzes data with respect to various aspects using (i) the amount of data transfers, (ii) destination IPs access patterns, (iii) employee's browsing patterns and (iv) employee's entry log into the restricted area. Any abnormality noticed is immediately reported to the user in the form of plots. In this presentation, we list out the various analyses performed and how they were utilized in the visualization. A few screenshots of the tool are provided to illustrate our analytic information presentation.

The IEEE Visual Analytics Science and Technology (VAST) Symposium has held a contest each year since its inception in 2006. These events are designed to provide visual analytics researchers and developers with analytic challenges similar to those encountered by professional information analysts. The VAST contest has had an extended life outside of the symposium, however, as materials are being used in universities and other educational settings, either to help teachers of visual analytics-related classes or for student projects. We describe how we develop VAST contest datasets that results in products that can be used in different settings and review some specific examples of the adoption of the VAST contest materials in the classroom. The examples are drawn from graduate and undergraduate courses at Virginia Tech and from the Visual Analytics ldquoSummer Camprdquo run by the National Visualization and Analytics Center in 2008. We finish with a brief discussion on evaluation metrics for education.

We hypothesized that potential spies would try to use other employees' terminals in order to not draw attention to themselves. We define one type of suspicious activity as IP use on a terminal when the owner is inside the classified area. We created a timeline visualization of IP usage, overlaid with classified area entrances and exits. The vertical axis divides the timelines into 31 rows, one for each day of the month. The horizontal axis represents the time of day from early morning to late evening. A single employee's entire month is viewed all at once using this visualization. The employee being viewed can be changed using the arrow keys. Every IP event is represented by a vertical bar positioned at the exact time of its appearance. We color the IP events by port number, which is either intranet, HTTP, tomcat, or email, and size the bar based on the outgoing data size. Whenever an employee enters the classified area, a semi-transparent yellow region is drawn until that user exits the classified area. In rare cases when the user double enters, the region is twice as opaque, and in the other rare case where a user leaves the exits without entering, a red region is drawn until the next time the employee enters. The legend key and office diagram showing the current selected employee, highlighted in red, can be seen in the top left-hand corner.

Today's challenging task in intelligent data processing is not to store large volumes of interlinked data but to visualize, explore, and understand its explicit or implicit relationships. Our solution to this is the VIScover system. VIScover combines semantic technologies with interactive exploration and visualization techniques able to analyze large volumes of structured data. We briefly describe our VIScover system and show its potential using the example of the VAST 2009 social network and geospatial data set.

In this paper, we present a system for the interactive visualization and exploration of graphs with many weakly connected components. The visualization of large graphs has recently received much research attention. However, specific systems for visual analysis of graph data sets consisting of many components are rare. In our approach, we rely on graph clustering using an extensive set of topology descriptors. Specifically, we use the self-organizing-map algorithm in conjunction with a user-adaptable combination of graph features for clustering of graphs. It offers insight into the overall structure of the data set. The clustering output is presented in a grid containing clusters of the connected components of the input graph. Interactive feature selection and task-tailored data views allow the exploration of the whole graph space. The system provides also tools for assessment and display of cluster quality. We demonstrate the usefulness of our system by application to a shareholder network analysis problem based on a large real-world data set. While so far our approach is applied to weighted directed graphs only, it can be used for various graph types.

Complex scenario analysis requires the exploration of multiple hypotheses and supporting evidence for each argument posed. Knowledge-intensive organisations typically analyse large amounts of inter-related, heterogeneous data to retrieve the knowledge this contains and use it to support effective decision-making. We demonstrate the use of interactive graph visualisation to support hierarchical, task-driven, hypothesis investigation. The visual investigative analysis is guided by task and domain ontologies used to capture the structure of the investigation process and the experience gained and knowledge created in previous, related investigations.